|
On October 12 2010 04:54 Empyrean wrote:
If you're looking for Bayesian Statistics, CMU and Duke have the strongest departments in the world.
Of course, obligatory Stanford, Cal, etc.
Thanks. I almost forgot about Duke!!! Best advice so far.
Added Duke to my list of schools to send applications to. :D
(Rick Durrett is at Duke. w00t. He is not statistician though.)
I will research on CMU later.
On October 12 2010 04:57 sLiniss wrote:
Boston University
In the middle of the city, clubs, hot college chicks. AHMG I'm having a blast here.
Boston University does not even have a stand-alone statistics department.
|
17082 Posts
On October 12 2010 05:03 Sufficiency wrote:
(Rick Durrett is at Duke. w00t. He is not statistician though.)
Alan Glefand is at Duke. He and Smith wrote "Sampling-Based Approaches to Calculating Marginal Densities," which is pretty much the most cited Bayesian paper in existence :O
EDIT: Link to the paper: http://www.jstor.org/stable/2289776
And yeah, definitely check out CMU too. They have a great Bayesian stat department.
|
The only problem I see is that I have no interests in Bayesian statistics. I am willing to keep an open mind about it, but..... o well, that paper is only 13 pages long so I will give it a shot. :D
Unfortunately I am not too sure what I am interested in. The only thing that come to mind right now is analysis of dependent data.
|
UCLA's stats department has a poker tournament to kick off the year and I went to that, all the grad students and professors seemed happy to be there at least although I don't know the rankings. UCLA's probably a better choice for a math PhD because you'd get to take Terrence Tao's classes.
|
On October 12 2010 05:42 jalstar wrote:
UCLA's stats department has a poker tournament to kick off the year and I went to that, all the grad students and professors seemed happy to be there at least although I don't know the rankings. UCLA's probably a better choice for a math PhD because you'd get to take Terrence Tao's classes.
Assuming I get into UCLA's statistics, I am sure I can make up excuses to go to Terrence Tao's lectures anyways. I am very fluent in analysis and should be able to chew most of the things he throws at me.
E: UCLA's statistics faculties are quite sad, actually (at least looking at the list from the surface). But I suppose it's still worthwhile to throw one application in there.
|
Apply to the University of Michigan  from what I hear we have a strong stats dept.
|
Virginia Tech... I will be honest I don't know much about the stat department, I know we have one though unlike the BU suggestion. But great place to live pretty good stipends usually around 20k (and apartments around here are dirt cheap.) I would take a look at the program see if it is any good, and consider the school if it is because everything else is great.
|
On October 12 2010 06:28 AcrossFiveJulys wrote:Apply to the University of Michigan from what I hear we have a strong stats dept.
UM Ann Arbor is on my list 
Maybe I should write a list of universities I already have in mind. One minute...
On October 12 2010 06:34 jamesr12 wrote:
Virginia Tech... I will be honest I don't know much about the stat department, I know we have one though unlike the BU suggestion. But great place to live pretty good stipends usually around 20k (and apartments around here are dirt cheap.) I would take a look at the program see if it is any good, and consider the school if it is because everything else is great.
Researched. Thanks.
|
On October 12 2010 06:46 Sufficiency wrote:Show nested quote +On October 12 2010 06:28 AcrossFiveJulys wrote:Apply to the University of Michigan from what I hear we have a strong stats dept. UM Ann Arbor is on my list Maybe I should write a list of universities I already have in mind. One minute...
Ann Arbor is a whore, go to The Ohio State University!
We're ranked 20th among all graduate programs in statistics and ranked 12th among all public institutions with graduate programs in statistics (or so says the stats website) http://www.stat.osu.edu/ if you want to check it out
|
there seem to be plenty of good Bayesians at CMU
|
On October 12 2010 04:38 Glaven wrote:Isn't harvey mudd primarily an undergraduate college? Or am I mistaken?
You are correct:
http://en.wikipedia.org/wiki/Harvey_Mudd_College#Reputation
"In 1997, Harvey Mudd College became the sole American undergraduate-only institution ever to win 1st place in the ACM International Collegiate Programming Contest"
|
On October 12 2010 06:50 Molybdenum wrote:Show nested quote +On October 12 2010 06:46 Sufficiency wrote:On October 12 2010 06:28 AcrossFiveJulys wrote:Apply to the University of Michigan from what I hear we have a strong stats dept. UM Ann Arbor is on my list Maybe I should write a list of universities I already have in mind. One minute... Ann Arbor is a whore, go to The Ohio State University! We're ranked 20th among all graduate programs in statistics and ranked 12th among all public institutions with graduate programs in statistics (or so says the stats website) http://www.stat.osu.edu/ if you want to check it out
Researched. Thanks.
On October 12 2010 06:58 palanq wrote:
there seem to be plenty of good Bayesians at CMU
Hmmm... I think I will apple for one of Duke or CMU. Frankly Bayesian statistics is not what I truly like. I won't say I hate them, but...
|
17082 Posts
On October 12 2010 11:06 Sufficiency wrote:Show nested quote +On October 12 2010 06:50 Molybdenum wrote:On October 12 2010 06:46 Sufficiency wrote:On October 12 2010 06:28 AcrossFiveJulys wrote:Apply to the University of Michigan from what I hear we have a strong stats dept. UM Ann Arbor is on my list Maybe I should write a list of universities I already have in mind. One minute... Ann Arbor is a whore, go to The Ohio State University! We're ranked 20th among all graduate programs in statistics and ranked 12th among all public institutions with graduate programs in statistics (or so says the stats website) http://www.stat.osu.edu/ if you want to check it out Researched. Thanks. Show nested quote +On October 12 2010 06:58 palanq wrote:
there seem to be plenty of good Bayesians at CMU Hmmm... I think I will apple for one of Duke or CMU. Frankly Bayesian statistics is not what I truly like. I won't say I hate them, but...
Going to be honest here: computationally intensive methods are the future of statistics.
|
On October 12 2010 13:34 Empyrean wrote:Show nested quote +On October 12 2010 11:06 Sufficiency wrote:On October 12 2010 06:50 Molybdenum wrote:On October 12 2010 06:46 Sufficiency wrote:On October 12 2010 06:28 AcrossFiveJulys wrote:Apply to the University of Michigan from what I hear we have a strong stats dept. UM Ann Arbor is on my list Maybe I should write a list of universities I already have in mind. One minute... Ann Arbor is a whore, go to The Ohio State University! We're ranked 20th among all graduate programs in statistics and ranked 12th among all public institutions with graduate programs in statistics (or so says the stats website) http://www.stat.osu.edu/ if you want to check it out Researched. Thanks. On October 12 2010 06:58 palanq wrote:
there seem to be plenty of good Bayesians at CMU Hmmm... I think I will apple for one of Duke or CMU. Frankly Bayesian statistics is not what I truly like. I won't say I hate them, but... Going to be honest here: computationally intensive methods are the future of statistics.
Yea I know.... Ahh what the heck. I'll throw an application to CMU too.
Right now I have 14 schools listed. Is this a reasonable number of schools to apply to? (Application fee isn't a problem... I am more worried about my references complaining  )
|
On October 12 2010 13:39 Sufficiency wrote:Show nested quote +On October 12 2010 13:34 Empyrean wrote:On October 12 2010 11:06 Sufficiency wrote:On October 12 2010 06:50 Molybdenum wrote:On October 12 2010 06:46 Sufficiency wrote:On October 12 2010 06:28 AcrossFiveJulys wrote:Apply to the University of Michigan from what I hear we have a strong stats dept. UM Ann Arbor is on my list Maybe I should write a list of universities I already have in mind. One minute... Ann Arbor is a whore, go to The Ohio State University! We're ranked 20th among all graduate programs in statistics and ranked 12th among all public institutions with graduate programs in statistics (or so says the stats website) http://www.stat.osu.edu/ if you want to check it out Researched. Thanks. On October 12 2010 06:58 palanq wrote:
there seem to be plenty of good Bayesians at CMU Hmmm... I think I will apple for one of Duke or CMU. Frankly Bayesian statistics is not what I truly like. I won't say I hate them, but... Going to be honest here: computationally intensive methods are the future of statistics. Yea I know.... Ahh what the heck. I'll throw an application to CMU too. Right now I have 14 schools listed. Is this a reasonable number of schools to apply to? (Application fee isn't a problem... I am more worried about my references complaining )
So long as you're on good terms w/ your research advisor, shouldn't be a big deal.
Most of the stuff is copy-pasta'd anyways.
|
Georgia Tech has a pretty decent stats program and it has a great environment for math/computers/robotics/AI/etc.
don't under value the location of your school - where you'll be living and the general surroundings is hugely important, especially since a PhD program is a few years at least (right?)
|
I don't think you should rule out Canadian schools so quickly; although I am in the US, I have a very positive impression of schools such as UofT and UBC.
Further, just because a school is housed in a famous university (Harvard, Yale) does not mean it is an attractive choice. If adjacent departments such as EECS or mathematics are weak or outside of focus, it is actually a negative. Having the best algebraic geometers in the world is pointless, if you need to go to a nearby university to learn wavelets and harmonic analysis. So GIT isn't a bad suggestion.
The list is too long but you have a lot of places I don't recommend trying like Cornell, Columbia. It may sadden you but you're better off going to places in North Carolina, Connecticut, and Illinois, if you want to do research.
As a student now, you can go and ask about departments, and people will tell you: "oh that place is big, lots of resources, people doing different things," but once you start to figure out these things - trust me the strength of other departments is very local.
Honestly, only 2, arguably 3, departments are as uniformly strong (and the most difficult to gain admission to) as most departments believe they are.
If you're interest is in applied probability, you should consider applied mathematics and operations research; a good theoretical OR department can give you good training. Although it sounds like your interest is in probability?, so honestly within statistics itself it is difficult to find both very knowledgeable and very active in research probabilists (outside of places like Berkeley).
But yes, if your interest is in concentration of measure phenomena for dependence in things like statistical physics, you probably should be applying for math/applied math programs.
My advice is, given my interpretation of your post, unfortunately to figure out specifically what you want to do (theory, computation, probability, machine learning, games/decision) and sacrifice the rest. Check the genealogy project for students and placements; if you see a big genius with 50 publications in the Annals but no students there's probably a very good reason for this (personality, etc...).
Going to school like Duke is good, but you are giving up a lot. Duke is able to be competitive precisely because of the way the faculty are guiding the department. I'd say that most people going to Duke will end up doing computational (a lot more so than theory) Bayesian stuff. But to second Empyrean, Gelfand is an expert on spatially dependent data
|
On October 12 2010 05:47 Sufficiency wrote:Show nested quote +On October 12 2010 05:42 jalstar wrote:
UCLA's stats department has a poker tournament to kick off the year and I went to that, all the grad students and professors seemed happy to be there at least although I don't know the rankings. UCLA's probably a better choice for a math PhD because you'd get to take Terrence Tao's classes. Assuming I get into UCLA's statistics, I am sure I can make up excuses to go to Terrence Tao's lectures anyways. I am very fluent in analysis and should be able to chew most of the things he throws at me. E: UCLA's statistics faculties are quite sad, actually (at least looking at the list from the surface). But I suppose it's still worthwhile to throw one application in there.
I wouldn't call Mark Handcock sad, and also the editor of the J of Multivariate Analysis is there. If you ignore warning signs such as the number of students graduating, time to graduation, and happiness, you're inviting misery.
|
|
|
|
|
|
EPT
